Why University Endowments are Large and Risky

Thomas Gilbert is an Assistant Professor of Finance & Business Economics at the University of Washington. This post is based on an article authored by Professor Gilbert and Christopher Hrdlicka, Assistant Professor of Finance & Business Economics at the University of Washington.

Universities as perpetual ivory towers, though often meant as a pejorative, describes well universities’ special place in society as centers of learning with a mission distinct from that of businesses. Universities create new knowledge via research while preserving and spreading that knowledge through teaching. The social good aspect of universities makes donations critical to funding their mission. But rather than investing these donations internally to build the metaphorical towers higher and shine the light of learning more widely, universities have built large endowments invested heavily in risky financial assets.

In our paper, Why Are University Endowments Large and Risky?, forthcoming at The Review of Financial Studies, we model how universities’ objectives, investment opportunities (internal and external) and public policy, specifically the Uniform Prudent Management of Institutional Funds Act (UPMIFA), interact to create this behavior. Our findings suggest a reevaluation of UPMIFA’s ability to achieve its goal of maintaining donor intent in light of the costs it imposes on universities.

The average U.S. research and baccalaureate university has an endowment worth more than three times its annual budget. It invests this endowment predominantly in risky assets, with more than 75% of it in securities such as equities, hedge funds, real estate, private equity, and other alternative assets. Moreover, the average university’s donation flow is large, making up almost 20% of its annual budget, and actually even larger than its endowment payout, leading to a net endowment pay-in. While having an endowment makes sense since donations are volatile, these donations do rise and fall with financial markets. This means that a riskily invested endowment loses value at the same time as donations decrease. We saw just this in the 2008 financial crisis, and it led the most prominent universities with the largest endowments to forgo or reduce the production of research and teaching.

In our paper, we study the forces that drive universities (and nonprofits more generally) that have productive internal projects sufficient to motivate such large donations to tolerate the opportunity cost of building large and risky endowments. The basic intuition behind our model is that building a large endowment is not an end unto itself but instead comes with an opportunity cost. As producers of what we call social dividends (research and teaching), this opportunity cost is forgone internal investment that would increase the output of social dividends. Universities also face the subtler but equally important trade-off that taking risk externally in financial markets limits the risk they can take in new internal projects.

We build a model to consider a range of university objectives that balance these trade-offs. At one extreme, the university’s decision makers seek only to maximize the production of social dividends. At the other extreme, the decision makers maximize the production of a nonpublic good that captures agency costs, conflicts of interest, perquisites present throughout the university structure, and even prestige maximization, such as caring about the quality of the average student or alumni. Since university production and its opportunity cost are the main features of our model, we consider variations in the productivity (quality) of the university’s internal projects as proxied by the amount of donations raised per unit of internal capital. The productivity of projects is inherently tied to the level of competition the university faces: more competition means higher costs and/or lower revenue (donations), i.e., lower productivity.

In the model where only social dividends are valued, the university chooses to forgo new internal investment in social dividends only if the expected return on this internal capital is lower than the expected return offered externally by the external financial markets. This base model is unable to generate the large and risky endowments observed in the data. Instead, the university takes risk by investing internally to expand production and uses the endowment primarily as a precautionary savings buffer. When the university has a low opportunity cost, that is, it generates a low level of donations per unit of capital (low productivity or low quality projects), it invests less internally and holds a large and safe endowment. As the opportunity cost increases, that is, as donations rise per unit capital (high productivity or high quality projects), the university invests more internally and holds a smaller and riskier endowment.

As we move to the other extreme of objectives where only the nonpublic good matters to the university’s decision makers, we find that the university invests less internally and builds a large and risky endowment. This occurs because the individuals in control, aware of the conflicts of interest within the university, realize that their claim to the donations gets diluted with any increase in internal capital. Adding this alternative objective produces endowment investment and payout decisions that more closely match the data, and it can even generate endowments that grow without bounds.

Finally, we model the influence of public policy, specifically UPMIFA, which, in order to maintain donor intent, mandates a maximum endowment payout constraint of 7% per year. We show that this maximum payout constraint has little effect on the second type of university, those that place little value on social dividends, since they do not invest internally even without the constraint. The constraint, however, has large effects on the first type of university, those that do value social dividends. By preventing large payouts in bad states, the constraint prevents the endowment’s effective use as precautionary buffer. Without the ability to use the endowment as a rainy day fund, these constrained universities choose instead to build large risky endowments consistent with the data. Moreover the payout constraint prevents universities from undertaking all of their productive internal projects since it also prevents large payouts in good states. The constraint thus harms these universities and society: increasing the volatility of social dividend production and slowing the growth of the most productive universities.

The full paper is available for download here.

Both comments and trackbacks are currently closed.